A5089854039- Underwater Acoustics Research
- Polyoxometalates: Synthesis and Applications
- Geophysical and Geoelectrical Methods
- Protein Interaction Studies and Fluorescence Analysis
- Geophysics and Gravity Measurements
- Geological and Geophysical Studies
- Image and Signal Denoising Methods
- Catalysis and Hydrodesulfurization Studies
- Underwater Vehicles and Communication Systems
- Drug Transport and Resistance Mechanisms
- Advanced Neural Network Applications
- Mesoporous Materials and Catalysis
- Advanced Photocatalysis Techniques
- Catalysis for Biomass Conversion
- Chemical Synthesis and Reactions
- Maritime and Coastal Archaeology
- Seismic Imaging and Inversion Techniques
- Advanced Image Processing Techniques
- Material Science and Thermodynamics
- Oceanographic and Atmospheric Processes
- Ferroelectric and Piezoelectric Materials
- Microwave Dielectric Ceramics Synthesis
- Maritime Navigation and Safety
- Luminescence Properties of Advanced Materials
- Glass properties and applications
Dalian Ocean University
2011-2025
Dalian Naval Academy
2015-2025
Chinese Academy of Sciences
2022-2025
Shanghai Institute of Ceramics
2022-2025
University of Chinese Academy of Sciences
2022-2025
Qingdao University
2024
Beijing Jiaotong University
2024
Hebei University
2013-2023
Baoding University
2017-2021
Commercial Aircraft Corporation of China (China)
2021
When used to recognize side-scan sonar images of shipwreck targets, the Faster R-CNN model is time-consuming and has low efficiency a high missed detection rate for small targets. Considering that existing datasets targets are small, we propose YOLOv3 can automatically based on transfer learning. Based Darknet-53 network, froze part convolutional layer trained COCO dataset images, conducted Multi-scale training shallow feature fusion was done multi-scale with Feature Pyramid Networks (FPN)...
Given the lack of systematic research on bathymetric surveys with multi-beam sonar carried by autonomous underwater vehicles (AUVs) in unfamiliar waters, this paper proposes a method for based constant-depth mode AUVs, considering equipment safety, operational efficiency, and data quality. Firstly, basic principles under are proposed standards AUV characteristics. Secondly, vertical effective height model vehicle is established, providing constraints basis determining fixed depth missions....
Catalytic hydrodeoxygenation of 5-hydroxymethylfurfural into 2,5-dimethylfuran has received great interest in recent years.
The scarcity and difficulty in acquiring Side-scan sonar target images limit the application of deep learning algorithms detection. At present, there are few amplification methods for images, image quality is not ideal, which suitable characteristics images. Addressing current shortage sample augmentation sonar, this paper proposes a method augmenting single underwater using CBL-sinGAN network. Firstly, considering low resolution monochromatic nature while balancing training efficiency...
Aiming at the problem of low accuracy multi-scale seafloor target detection in side-scan sonar images with high noise and complex background texture, a model for using BES-YOLO network is proposed. First, an efficient attention (EMA) mechanism used backbone YOLOv8 network, bi-directional feature pyramid (Bifpn) introduced to merge information different scales, finally, Shape_IoU loss function continuously optimize improve its accuracy. Before training, dataset preprocessed 2D discrete...
The characteristics of multi-directional rotation and high aspect ratio targets such as shipwrecks lead to low detection accuracy difficulty localizing existing models for this target type. Through our research, we design three main inconsistencies in rotating compared traditional detection, i.e., inconsistency between anchor frame, classification features regression features, frame quality label assignment strategy. In paper, address the discrepancies above aspects, propose Side-scan Sonar...
Due to the large number of parameters and high computational complexity current target detection models, it is challenging perform fast accurate in side-scan sonar images under existing technical conditions, especially environments with limited resources. Moreover, since original waterfall map only consists echo intensity information, which usually a size, difficult fuse other multi-source limits accuracy models. To address these issues, we designed DBnet, lightweight detector featuring two...
The calculation of phonon scattering coefficients is used to study the influence different structures on glass-like thermal conductivity.
Abstract Background: Colorectal cancer (CRC) is one of the most common cancers worldwide and requires more effective safer therapies to improve its poor survival outcome. LY6G6D a GPI-anchored cell surface protein with expression that particularly specific colorectal (CRC). exhibits significantly elevated levels in tissue comparing normal tissue, pMMR/MSS CRC, which demonstrates primary resistance immune checkpoint inhibitor (ICI) therapy. T engagers have shown encouraging efficacy treating...
In response to the current challenges in efficiently acquiring sound speed profiles and ensuring their representativeness, considering need fully leverage historical while accounting for spatiotemporal variability, we introduce a model profile prediction based on CNN-BiLSTM-Attention network, which integrates convolutional neural network (CNN), bidirectional long short-term memory (BiLSTM), an attention mechanism (AM). The synergy of these components enables extract features more...
Self-healable hydrogels with cross-linking induced thermo-responsiveness (CIT) were prepared from ketone-group containing P(DMA-<italic>stat</italic>-DAA).
A novel Schiff (Mo) base complex supported on GO/g-C<sub>3</sub>N<sub>4</sub> was prepared and applied to the photocatalytic epoxidation reactions.
Na<sub>4</sub>SnS<sub>4</sub> and Na<sub>4</sub>SnSe<sub>4</sub> exhibiting excellent physicochemical performances as potential IR NLO crystals sodium ion conductors were systematically studied.
Current methods of fusing side-scan sonar images fail to tackle the issues shadow removal, preservation information from adjacent strip images, and maintenance image clarity contrast. To address these deficiencies, a novel curvelet-transform-based approach that integrates complementary attribute details is proposed. By capitalizing on multiple scales orientations curvelet transform its intricate hierarchical nature, myriad fusion rules were applied at corresponding frequency levels, enabling...